The Population Council and CONASIDA (the National AIDS committee of Mexico) will work together to gather and analyze behavioral data which can contribute to the design of programs to reduce HIV transmission among bisexuals in Mexico. Building on a previous cooperative research relationship between the two institutions, the project seeks to strengthen the capacity for in-depth behavioral research, training and intervention on AIDS in Mexico and Latin America. A rapid increase in Mexico of heterosexual transmission of HIV due to sexual intercourse has led to concern with male bisexuals as the transmission link between high prevalence populations (homosexuals) and currently low prevalence groups (women and children) in Mexico. Cultural factors including sanctions against homosexual behavior, the lack of a gay culture, and pressure to enter into a traditional marriage may lead to under reporting and under-estimation of the size of the bisexual population in Mexico. In Phase I, the research will consist of individual semi-structured interviews designed to elicit information regarding bisexual men. Of specific interest are high risk practices, partner networks, attitudes toward partner protection, strategies for dealing with HIV risk and amenability to change. The project will also study a group of women partners of bisexual men, in order to understand how cultural and socioeconomic circumstances may affect their ability to negotiate protection from HIV. In Phase II, two parallel activities will be undertaken; a survey to estimate the size and characteristics of the bisexual population in Mexico City, and three pilot interventions based on the social influence model. Intervention sites will include the university, an STD clinic and a community based, "storefront" drop-in center. At each site, project members will provide information, but the perceived social influence operating to induce behavior change will vary. Degrees of participation in the program (dosage) and reported change in high-risk practices will be outcome measures. The use of appropriate control groups and a longitudinal design will allow the measure of the impact of the intervention.